1 //===- InputFiles.cpp -----------------------------------------------------===//
2 //
3 //                             The LLVM Linker
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 
10 #include "Chunks.h"
11 #include "Error.h"
12 #include "InputFiles.h"
13 #include "Symbols.h"
14 #include "llvm/ADT/STLExtras.h"
15 #include "llvm/IR/LLVMContext.h"
16 #include "llvm/LTO/LTOModule.h"
17 #include "llvm/Object/COFF.h"
18 #include "llvm/Support/COFF.h"
19 #include "llvm/Support/Debug.h"
20 #include "llvm/Support/Endian.h"
21 #include "llvm/Support/raw_ostream.h"
22 
23 using namespace llvm::COFF;
24 using namespace llvm::object;
25 using namespace llvm::support::endian;
26 using llvm::Triple;
27 using llvm::support::ulittle32_t;
28 using llvm::sys::fs::file_magic;
29 using llvm::sys::fs::identify_magic;
30 
31 namespace lld {
32 namespace coff {
33 
34 int InputFile::NextIndex = 0;
35 llvm::LLVMContext BitcodeFile::Context;
36 
37 // Returns the last element of a path, which is supposed to be a filename.
38 static StringRef getBasename(StringRef Path) {
39   size_t Pos = Path.find_last_of("\\/");
40   if (Pos == StringRef::npos)
41     return Path;
42   return Path.substr(Pos + 1);
43 }
44 
45 // Returns a string in the format of "foo.obj" or "foo.obj(bar.lib)".
46 std::string InputFile::getShortName() {
47   if (ParentName == "")
48     return getName().lower();
49   std::string Res = (getBasename(ParentName) + "(" +
50                      getBasename(getName()) + ")").str();
51   return StringRef(Res).lower();
52 }
53 
54 void ArchiveFile::parse() {
55   // Parse a MemoryBufferRef as an archive file.
56   auto ArchiveOrErr = Archive::create(MB);
57   error(ArchiveOrErr, "Failed to parse static library");
58   File = std::move(*ArchiveOrErr);
59 
60   // Allocate a buffer for Lazy objects.
61   size_t NumSyms = File->getNumberOfSymbols();
62   LazySymbols.reserve(NumSyms);
63 
64   // Read the symbol table to construct Lazy objects.
65   for (const Archive::Symbol &Sym : File->symbols())
66     LazySymbols.emplace_back(this, Sym);
67 
68   // Seen is a map from member files to boolean values. Initially
69   // all members are mapped to false, which indicates all these files
70   // are not read yet.
71   for (auto &ChildOrErr : File->children()) {
72     error(ChildOrErr, "Failed to parse static library");
73     const Archive::Child &Child = *ChildOrErr;
74     Seen[Child.getChildOffset()].clear();
75   }
76 }
77 
78 // Returns a buffer pointing to a member file containing a given symbol.
79 // This function is thread-safe.
80 MemoryBufferRef ArchiveFile::getMember(const Archive::Symbol *Sym) {
81   auto COrErr = Sym->getMember();
82   error(COrErr, Twine("Could not get the member for symbol ") + Sym->getName());
83   const Archive::Child &C = *COrErr;
84 
85   // Return an empty buffer if we have already returned the same buffer.
86   if (Seen[C.getChildOffset()].test_and_set())
87     return MemoryBufferRef();
88   ErrorOr<MemoryBufferRef> Ret = C.getMemoryBufferRef();
89   error(Ret, Twine("Could not get the buffer for the member defining symbol ") +
90                  Sym->getName());
91   return *Ret;
92 }
93 
94 void ObjectFile::parse() {
95   // Parse a memory buffer as a COFF file.
96   auto BinOrErr = createBinary(MB);
97   if (!BinOrErr)
98     error(errorToErrorCode(BinOrErr.takeError()),
99                            "Failed to parse object file");
100   std::unique_ptr<Binary> Bin = std::move(*BinOrErr);
101 
102   if (auto *Obj = dyn_cast<COFFObjectFile>(Bin.get())) {
103     Bin.release();
104     COFFObj.reset(Obj);
105   } else {
106     error(Twine(getName()) + " is not a COFF file.");
107   }
108 
109   // Read section and symbol tables.
110   initializeChunks();
111   initializeSymbols();
112   initializeSEH();
113 }
114 
115 void ObjectFile::initializeChunks() {
116   uint32_t NumSections = COFFObj->getNumberOfSections();
117   Chunks.reserve(NumSections);
118   SparseChunks.resize(NumSections + 1);
119   for (uint32_t I = 1; I < NumSections + 1; ++I) {
120     const coff_section *Sec;
121     StringRef Name;
122     std::error_code EC = COFFObj->getSection(I, Sec);
123     error(EC, Twine("getSection failed: #") + Twine(I));
124     EC = COFFObj->getSectionName(Sec, Name);
125     error(EC, Twine("getSectionName failed: #") + Twine(I));
126     if (Name == ".sxdata") {
127       SXData = Sec;
128       continue;
129     }
130     if (Name == ".drectve") {
131       ArrayRef<uint8_t> Data;
132       COFFObj->getSectionContents(Sec, Data);
133       Directives = std::string((const char *)Data.data(), Data.size());
134       continue;
135     }
136     // Skip non-DWARF debug info. MSVC linker converts the sections into
137     // a PDB file, but we don't support that.
138     if (Name == ".debug" || Name.startswith(".debug$"))
139       continue;
140     // We want to preserve DWARF debug sections only when /debug is on.
141     if (!Config->Debug && Name.startswith(".debug"))
142       continue;
143     if (Sec->Characteristics & llvm::COFF::IMAGE_SCN_LNK_REMOVE)
144       continue;
145     auto *C = new (Alloc) SectionChunk(this, Sec);
146     Chunks.push_back(C);
147     SparseChunks[I] = C;
148   }
149 }
150 
151 void ObjectFile::initializeSymbols() {
152   uint32_t NumSymbols = COFFObj->getNumberOfSymbols();
153   SymbolBodies.reserve(NumSymbols);
154   SparseSymbolBodies.resize(NumSymbols);
155   llvm::SmallVector<std::pair<Undefined *, uint32_t>, 8> WeakAliases;
156   int32_t LastSectionNumber = 0;
157   for (uint32_t I = 0; I < NumSymbols; ++I) {
158     // Get a COFFSymbolRef object.
159     auto SymOrErr = COFFObj->getSymbol(I);
160     error(SymOrErr, Twine("broken object file: ") + getName());
161 
162     COFFSymbolRef Sym = *SymOrErr;
163 
164     const void *AuxP = nullptr;
165     if (Sym.getNumberOfAuxSymbols())
166       AuxP = COFFObj->getSymbol(I + 1)->getRawPtr();
167     bool IsFirst = (LastSectionNumber != Sym.getSectionNumber());
168 
169     SymbolBody *Body = nullptr;
170     if (Sym.isUndefined()) {
171       Body = createUndefined(Sym);
172     } else if (Sym.isWeakExternal()) {
173       Body = createUndefined(Sym);
174       uint32_t TagIndex =
175           static_cast<const coff_aux_weak_external *>(AuxP)->TagIndex;
176       WeakAliases.emplace_back((Undefined *)Body, TagIndex);
177     } else {
178       Body = createDefined(Sym, AuxP, IsFirst);
179     }
180     if (Body) {
181       SymbolBodies.push_back(Body);
182       SparseSymbolBodies[I] = Body;
183     }
184     I += Sym.getNumberOfAuxSymbols();
185     LastSectionNumber = Sym.getSectionNumber();
186   }
187   for (auto WeakAlias : WeakAliases)
188     WeakAlias.first->WeakAlias = SparseSymbolBodies[WeakAlias.second];
189 }
190 
191 Undefined *ObjectFile::createUndefined(COFFSymbolRef Sym) {
192   StringRef Name;
193   COFFObj->getSymbolName(Sym, Name);
194   return new (Alloc) Undefined(Name);
195 }
196 
197 Defined *ObjectFile::createDefined(COFFSymbolRef Sym, const void *AuxP,
198                                    bool IsFirst) {
199   StringRef Name;
200   if (Sym.isCommon()) {
201     auto *C = new (Alloc) CommonChunk(Sym);
202     Chunks.push_back(C);
203     return new (Alloc) DefinedCommon(this, Sym, C);
204   }
205   if (Sym.isAbsolute()) {
206     COFFObj->getSymbolName(Sym, Name);
207     // Skip special symbols.
208     if (Name == "@comp.id")
209       return nullptr;
210     // COFF spec 5.10.1. The .sxdata section.
211     if (Name == "@feat.00") {
212       if (Sym.getValue() & 1)
213         SEHCompat = true;
214       return nullptr;
215     }
216     return new (Alloc) DefinedAbsolute(Name, Sym);
217   }
218   int32_t SectionNumber = Sym.getSectionNumber();
219   if (SectionNumber == llvm::COFF::IMAGE_SYM_DEBUG)
220     return nullptr;
221 
222   // Reserved sections numbers don't have contents.
223   if (llvm::COFF::isReservedSectionNumber(SectionNumber))
224     error(Twine("broken object file: ") + getName());
225 
226   // This symbol references a section which is not present in the section
227   // header.
228   if ((uint32_t)SectionNumber >= SparseChunks.size())
229     error(Twine("broken object file: ") + getName());
230 
231   // Nothing else to do without a section chunk.
232   auto *SC = cast_or_null<SectionChunk>(SparseChunks[SectionNumber]);
233   if (!SC)
234     return nullptr;
235 
236   // Handle section definitions
237   if (IsFirst && AuxP) {
238     auto *Aux = reinterpret_cast<const coff_aux_section_definition *>(AuxP);
239     if (Aux->Selection == IMAGE_COMDAT_SELECT_ASSOCIATIVE)
240       if (auto *ParentSC = cast_or_null<SectionChunk>(
241               SparseChunks[Aux->getNumber(Sym.isBigObj())]))
242         ParentSC->addAssociative(SC);
243     SC->Checksum = Aux->CheckSum;
244   }
245 
246   auto *B = new (Alloc) DefinedRegular(this, Sym, SC);
247   if (SC->isCOMDAT() && Sym.getValue() == 0 && !AuxP)
248     SC->setSymbol(B);
249 
250   return B;
251 }
252 
253 void ObjectFile::initializeSEH() {
254   if (!SEHCompat || !SXData)
255     return;
256   ArrayRef<uint8_t> A;
257   COFFObj->getSectionContents(SXData, A);
258   if (A.size() % 4 != 0)
259     error(".sxdata must be an array of symbol table indices");
260   auto *I = reinterpret_cast<const ulittle32_t *>(A.data());
261   auto *E = reinterpret_cast<const ulittle32_t *>(A.data() + A.size());
262   for (; I != E; ++I)
263     SEHandlers.insert(SparseSymbolBodies[*I]);
264 }
265 
266 MachineTypes ObjectFile::getMachineType() {
267   if (COFFObj)
268     return static_cast<MachineTypes>(COFFObj->getMachine());
269   return IMAGE_FILE_MACHINE_UNKNOWN;
270 }
271 
272 StringRef ltrim1(StringRef S, const char *Chars) {
273   if (!S.empty() && strchr(Chars, S[0]))
274     return S.substr(1);
275   return S;
276 }
277 
278 void ImportFile::parse() {
279   const char *Buf = MB.getBufferStart();
280   const char *End = MB.getBufferEnd();
281   const auto *Hdr = reinterpret_cast<const coff_import_header *>(Buf);
282 
283   // Check if the total size is valid.
284   if ((size_t)(End - Buf) != (sizeof(*Hdr) + Hdr->SizeOfData))
285     error("broken import library");
286 
287   // Read names and create an __imp_ symbol.
288   StringRef Name = StringAlloc.save(StringRef(Buf + sizeof(*Hdr)));
289   StringRef ImpName = StringAlloc.save(Twine("__imp_") + Name);
290   const char *NameStart = Buf + sizeof(coff_import_header) + Name.size() + 1;
291   DLLName = StringRef(NameStart);
292   StringRef ExtName;
293   switch (Hdr->getNameType()) {
294   case IMPORT_ORDINAL:
295     ExtName = "";
296     break;
297   case IMPORT_NAME:
298     ExtName = Name;
299     break;
300   case IMPORT_NAME_NOPREFIX:
301     ExtName = ltrim1(Name, "?@_");
302     break;
303   case IMPORT_NAME_UNDECORATE:
304     ExtName = ltrim1(Name, "?@_");
305     ExtName = ExtName.substr(0, ExtName.find('@'));
306     break;
307   }
308   ImpSym = new (Alloc) DefinedImportData(DLLName, ImpName, ExtName, Hdr);
309   SymbolBodies.push_back(ImpSym);
310 
311   // If type is function, we need to create a thunk which jump to an
312   // address pointed by the __imp_ symbol. (This allows you to call
313   // DLL functions just like regular non-DLL functions.)
314   if (Hdr->getType() != llvm::COFF::IMPORT_CODE)
315     return;
316   ThunkSym = new (Alloc) DefinedImportThunk(Name, ImpSym, Hdr->Machine);
317   SymbolBodies.push_back(ThunkSym);
318 }
319 
320 void BitcodeFile::parse() {
321   // Usually parse() is thread-safe, but bitcode file is an exception.
322   std::lock_guard<std::mutex> Lock(Mu);
323 
324   Context.ensureDITypeMap();
325   ErrorOr<std::unique_ptr<LTOModule>> ModOrErr = LTOModule::createFromBuffer(
326       Context, MB.getBufferStart(), MB.getBufferSize(), llvm::TargetOptions());
327   error(ModOrErr, "Could not create lto module");
328   M = std::move(*ModOrErr);
329 
330   llvm::StringSaver Saver(Alloc);
331   for (unsigned I = 0, E = M->getSymbolCount(); I != E; ++I) {
332     lto_symbol_attributes Attrs = M->getSymbolAttributes(I);
333     if ((Attrs & LTO_SYMBOL_SCOPE_MASK) == LTO_SYMBOL_SCOPE_INTERNAL)
334       continue;
335 
336     StringRef SymName = Saver.save(M->getSymbolName(I));
337     int SymbolDef = Attrs & LTO_SYMBOL_DEFINITION_MASK;
338     if (SymbolDef == LTO_SYMBOL_DEFINITION_UNDEFINED) {
339       SymbolBodies.push_back(new (Alloc) Undefined(SymName));
340     } else {
341       bool Replaceable =
342           (SymbolDef == LTO_SYMBOL_DEFINITION_TENTATIVE || // common
343            (Attrs & LTO_SYMBOL_COMDAT) ||                  // comdat
344            (SymbolDef == LTO_SYMBOL_DEFINITION_WEAK &&     // weak external
345             (Attrs & LTO_SYMBOL_ALIAS)));
346       SymbolBodies.push_back(new (Alloc) DefinedBitcode(this, SymName,
347                                                         Replaceable));
348     }
349   }
350 
351   Directives = M->getLinkerOpts();
352 }
353 
354 MachineTypes BitcodeFile::getMachineType() {
355   if (!M)
356     return IMAGE_FILE_MACHINE_UNKNOWN;
357   switch (Triple(M->getTargetTriple()).getArch()) {
358   case Triple::x86_64:
359     return AMD64;
360   case Triple::x86:
361     return I386;
362   case Triple::arm:
363     return ARMNT;
364   default:
365     return IMAGE_FILE_MACHINE_UNKNOWN;
366   }
367 }
368 
369 std::mutex BitcodeFile::Mu;
370 
371 } // namespace coff
372 } // namespace lld
373